Skip to main content

Full text of "Derailment and subsequent burning of Delaware and Hudson Railway freight train at Oneonta, New York, February 12, 1974"

See other formats



For Reference 
; Do Mot Take 
I From tlie 










FEBRUARY 12. 1974 

Washington, D.C. 20591 

^-^Tcuu ff- / / ou 








FEBRUARY 12, 1974 



Washington, D.C. 20591 


2. Government Accession No. 


1 . Report No. 

3. Recipient 's Catalog No. 

k. Title and Subtitle 

Railroad Accident Report - Derailment and subsequent 
burning of Delaware and Hudson Railway freight train 
at Oneonta, New York, February 12, 1974. 

5. Report Date 
October 17, 1974 

6. Performing Organization 

7. Author(s) 

8. Performing Organization 
Report No. 

9. Performing Organization Name and Address 

National Transportation Safety Board 
Bureau of Surface Transportation Safety 
Washington, D. C. 20591 

10. Work Unit No. 


11. Contract or Grant No. 

12. Sponsoring Agency Name and Address 

Washington, D. C. 20591 

13.Type of Report and 
Period Covered 
Railroad Accident Report 
February 12, 1974 

l4.Sponsori ng Agency Code 

IS-Supplementary Notes 

This report contains Railroad Safety Recommendations R-74-31 through R-74-34. 

16. Abstract 

This report describes and analyzes a derailment which occurred when a 
train separation resulted in unequal deceleration of the two parts of the 
train. Unusual lateral forces at the rear of the third locomotive unit 
canted the outside rail of a 3° 30' curve outward enough to allow the 
wheels to drop inside. A tank car of propane was punctured and the ensuing 
fire impinged other tank cars and caused the violent rupture of three of 
them. Fifty-four person were injured by the fire and rocketing parts of 
tank cars. 

The National Transportation Safety Board determines that the probable 
cause of this accident was the inability of the track to resist the lateral 
forces which canted the outside rail outward and widened the gage of the 
track. These forces which were induced at the third locomotive unit 
resulted from the emergency application of the brakes when the train was 
separated between the third and fourth cars as it entered the 3° 30' 
curve. The train separated as a result of the broken center sill on the 
fourth car. 

18. Distribution Statement 

This document is available 
to the public through the 
National Technical Informa- 
tion Service, Springfield, 
Virginia 22151 

17. Key Words 

Railroad accident. Derailment, Freight train. Broken 
Center Sill, Emergency Brake Application, Spread Track, 
Tank Car Repairs, Hazardous Material, Tank Car Ruptures 
Liquefied Petroleum Gas, Rocketing of Tank Cars. 

19. Security Classification 
(of this report) 

20. Security Classification 
(of this page) 

21 .No. of Pages 

22. Price 

NTSB Form 1765.2 (Rev. 9/74) 


The accident described in this report has been designated as a 
major accident by the National Transportation Safety Board under the 
criteria established in the Safety Board's regulations. 

This report is based on facts obtained from an investigation con- 
ducted by the Safety Board, in cooperation with the Federal Railroad 
Administration, The conclusions, the determination of probable cause, 
and the recommendations are those of the Safety Board. 






The Accident 1 

Location of Accident and Method of Operation 2 

The Track 4 

The Train 6 


The Accident 10 

Need for Inspections 11 

Track Structure 12 

Tank Cars 14 

Emergency Response by Firemen 16 

Communications between the Fire Departments and the 

Railroads 16 





WASHINGTON, D. C. 20591 


Adopted; October 17, 1974 

FEBRUARY 12, 1974 


About 4:20 p.m. on February 12, 1974, Delaware and Hudson Railway 
freight train NWB-4, derailed as it moved northward around a 3° 30' curve 
just north of Oneonta, New York. Of the derailed cars, seven were tank 
cars loaded with liquefied petroleum gas (LPG) . The tank of one of the 
cars split open in the derailment, and the escaping LPG ignited. Four of 
the remaining tank cars ruptured about 30 minutes after the derailment. 
The ensuing explosions and fire injured 54 firemen and members of the 

The National Transportation Safety Board determines that the prob- 
able cause of this accident was the inability of the track to resist the 
lateral forces which canted the outside rail outward and widened the 
gage of the track. These forces which were induced at the third loco- 
motive unit resulted from the emergency application of the brakes when 
the train was separated between the third and fourth cars as it entered 
the 3° 30' curve. The train separated as a result of the broken center 
sill on the fourth car. 


The Accident 

At 1:55 p.m. on February 12, 1974, Delaware and Hudson Railway 
Company's (D&H) freight train NWB-4 departed Binghamton, N.Y. , bound 
for Mechanicville, N, Y. The train consisted of 3 locomotive units and 
122 cars. Cars 20 through 27 were tank cars loaded with liquefied pet- 
roleum gas (LPG) . 

Inspections and brake tests were performed before the train departed 
Binghamton. The crewmembers made numerous inspections of the train en 
route. The train was slowed to 20 mph as it passed through Oneonta, 
N.Y. , where a carman inspected it as it passed. None of the inspections 
disclosed any defects. 

When NWB-4 was about 4 miles north of Oneonta, N.Y. , it was operating 
on the No. 1 track, the westermost track of the double-track section; its 

- 2 - 

speed was 32 mph. At 4:19 p.m. as NWB-4 rounded a 3° 30' curve, the 
crewmeiribers on the locomotive felt an unusual movement, looked toward the 
rear of the train, and saw some of the head cars rock violently and Car 
4 turn over on its side along the west side of the track. The fireman 
and brakeman instructed the engineer to apply the emergency brakes. The 
engineer applied the brakes immediately, but did not know if the braking 
was the result of his application or the result of the train's action. 
The front of the train moved about 1,200 feet and stopped. When the 
brakes were applied, the flagman in the caboose was thrown against the 
handrail and injured. 

Cars 5, 6, and 7 passed Car 4 and stopped in various positions just 
north of Car 4. (See Figure 1.) Cars 1, 2, and 3 remained coupled to 
the locomotive, but overturned when they stopped. 

Car 21, a tank car, was crushed by the derailing cars and split open, 
and escaping LPG was ignited immediately. The derailed cars were, in 
turn, set on fire. 

The engineer notified the railroad's Oneonta office by radio of the 
derailment and fire. The two front locomotive units were uncoupled from 
the train and moved northward about 1/2 mile. About 10 minutes after the 
derailment, firemen from the Oneonta Fire Department arrived and began to 
fight the fire. They were followed shortly by other companies. 

The intensity of the fire increased for 30 minutes and the tanks' 
sheets adjacent to the vapor space were heated to a point that they be- 
came thin, and the tanks consequently ruptured. As a result one of the 
cars exploded violently and several persons including firemen were in- 
jured. Three additional explosions occurred in 10-minute intervals fol- 
lowing the first. During the explosions, half sections of Cars 23 and 
25 were propelled eastward about 1,200 feet. Cars 22 and 24 separated 
into two tub-shaped ends, and portions split longitudinally. A small 
split occurred in the tank of Car 26 near the manway during the derail- 
ment. The escaping LPG ignited, was allowed to burn, and was consumed 
in about 7 days . 

The injured firemen were removed from the scene, and the others were 
instructed to pull back a safe distance. Car 27, which was not damaged, 
was rerailed and moved southward. 

Some typical damage outside the immediate area 3/4 to 1 mile from 
the site was as follows: Windows were broken, siding was blown from the 
side of a house, and a sink was blown from the wall of a house. 

Location of Accident and Method of Operation 

South of the 3° 30' curve, the tracks are straight for a considerable 
distance and the grade is 0.87 percent, ascending northward. A railroad 

- 3 


1st CB & O 20134—1 

2nd B 8, O 603034 

3rd C & O 602333 

12th NYC 892113 
18th PLE 32532 
21st UTLX 99489 
22nd UTLX 89268 

24th UTLX 83635 

26th GATX 96504 

27rti GATX 96536 

7th N & W 172192 

8th EFCX 1069 
9th IC 10348 

11th PTLX 41968 


20th CEI 119310 

23rd UTLX 83736 
25th GATX 96525 

Figure 1. Relative positions of derailed cars before explosions. 

- 4 - 

service road parallels the tracks on the east. Access roads were provided 
on each side of the accident site. (See Figure 2.) 

Trains are operated on both tracks in each direction by signals from 
a traffic control system which is controlled by a train dispatcher at 
Colonie, New York. The maximum authorized speed is 40 mph. 

The weather for the Oneonta vicinity on February 12 was : Low tem- 
perature - 9°F., high temperature - 35°F., cloudy, and 0.1 inch of snow 
had fallen. 

The Track 

Structure . The No. 1 track consisted of 112- lb. RE rails manufactured 
by the Bethlehem Steel Company in 1942, and laid in track the same year. 
The 39-foot rails were connected by 6-hole, head-free joint bars and were 
laid on 24 hardwood crossties per rail, with single and double shoulder 
tie plates intermixed. Originally, the D&H used two spikes per tie plate 
on a curve of 3° 30', but it had currently been using three spikes per tie 
plate as ties were replaced. However, there were no records to indicate 
the number of ties through the curve that had been secured with three 
spikes per tie plate. The rails were anchored with 16 to 24 rail anchors 
per rail length. The track was laid on crushed stone ballast, with full 
ballast sections. 

Inspection . The No. 1 track in the accident area was last inspected 
by a track foreman- inspector on February 11, 1974. His inspection was 
made on foot, and the only discrepancy noticed was that several joint 
bolts were loose. 

The Federal Railroad Administration's Track Inspector had inspected 
the tracks in the derailment area in October 1973. There were no dis- 
crepancies reported for Tracks 1 and 2 around the curve. 

Damage . About 960 feet of the No. 1 track was destroyed during the 
accident. The east, or high rail, was canted outward and intermittent 
wheel marks were found on the web of the rail from a point about 560 feet 
north of the south end of the curve to where the third locomotive unit 
stopped. The west rail was displaced westward by the wheels of the loco- 
motive unit and cars which dropped inside of the rail when the east rail 
moved outward. 

Since the crossties in the derailment area were badly damaged and 
burned, the number of spikes holding the rail and the condition of the 
crossties at the time of the accident could not be determined. 

Derailing cars destroyed 400 feet of the No. 2 track. 

- 5 - 

FEBRUARY 12, 1974 

15.6 MILES 
45.3 MILES 
3.9 MILES 
0.8 MILES 
0.6 MILES 


Figure 2. Area of the derailment. 

- 6 

The Train 

Locomotives . The lead locomotive was a 3,300 hp, type U33c, and the 
other two locomotives were 3,000 hp, type U30c. The locomotives were 
manufactured by the General Electric Company and were equipped with six 
wheel trucks. Each unit weighed about 390,000 pounds. 

Before the train was dispatched from Binghamton, the locomotives were 
inspected and tested. No discrepancies were noted. After the accident, 
the locomotives were again inspected, and no conditions were found that 
would have contributed to the cause of the accident. 

The wheels of the third locomotive were examined. Scrape marks were 
discovered on the back and front faces of the rims, and gouge marks were 
found on the flanges and treads which indicated that all of the wheels had 
derailed. When the locomotive was examined at the scene, all of the 
wheels on the west side were derailed inside the west rail. The wheels on 
the left side, except for the outside pair of wheels on the north truck, 
were riding on the web of the canted east rail. The outside pair of 
wheels of the north truck was on the rail. 

None of the wheels of the other units were derailed, and they did 
not bear any indications of having been derailed. 

Speed Recording Tapes . Calibration of the speed recorders on the 
three locomotive units disclosed that they accurately recorded the train's 
speed. The tapes were examined, and they indicated that the train was 
moving at 32 mph when the emergency brakes were applied. Each of the 
tapes contained evidence that the train had been subjected to a severe 
run-in when the emergency brakes were applied, and that the brakes were 
applied when the locomotive entered the 3° 30' curve. 

Car Equipment . Car 4, C&O 603325, was a covered hopper loaded with 
190,000 lbs, of shelled corn. The car had a capacity of 100 tons, a light- 
weight of 61,900 lbs., and a load unit of 201,100 lbs. The car was con- 
structed of welded steel and provided with 4 wheel-roller bearing trucks 
which had 36-inch wheels. The car was manufactured in December 1969 by 
the Pullman Standard Company. The car was provided with a continuous 
center sill, of AAR design, fabricated from two Z angles welded together 
with reinforcing angles. 

An examination of Car 4 disclosed that the center sill on the north 
end (B end) of the car was distorted upward about 7 1/2 inches. Marks 
were found on the "B" end coupler and on the knuckle of the coupler of 
Car 3, which indicated that the couplers had moved vertically one above 
the other and had become disengaged. This caused the train to separate. 
The center sill was cracked through the vertical web on the west side of 
the sill, 11 feet 4 inches from the north end where the center sill passed 

7 - 

between the hoppers of the car. The sill was cracked where the bottom 
flange of the center sill had been removed to permit application of the 
hopper. Rivets attached a reinforcing angle to the inside of the sill in 
this area. The progressive crack had started in the rivet hole and ex- 
tended through the remaining portion of the bottom flange, through the 
vertical web, and partially across the top flange. The sill was slightly 
bent upward at this point. The portion of the center sill which contained 
the fracture was removed and examined further. (See Figure 3.) 

Since the accident, the owner of the car, the Chessie System, 
selected randomly 15 percent of the covered hoppers, built to the same 
specifications, and found no indications of failure in the center sills. 
There are no reports of comparable failures in covered hoppers of similar 

An examination of the following cars disclosed that they too con- 
tained evidence that a heavy lateral force had been exerted in the train. 

Car 3 C&O 602333 - Striking casting on north end, east side, was 
broken. Coupler cross key was bent. 

Car 5 B&O 482721 - Marks on sides of striking casting and couplers 
indicated heavy contact. Striking casting on 
south end was bent. 

Car 6 PC 878941 - Coupler carrier iron was broken on one end and 
marks on striking casting and coupler indicated 
heavy contact. 

Car 7 N&W 172192 - Side of center sill was bent and broken on 
south end . 

The Tank Cars . Cars 20 through 27 were "jumbo-type" tank cars, and 
were loaded with an average of 30,200 gallons of LPG. These cars did not 
have continuous center sills, but had stub sills welded to the bottom of 
the tank at each end. The tanks were not insulated and were constructed 
to DOT Specification 112A-340W. An examination of the fractured surfaces 
of the tanks disclosed that some steel sheets had thinned to 3/8 inch. 
Their original thickness was 3/4 inch. 

Fire and Rescue . The Oneonta Fire Department was first notified of 
the derailment by the yardmaster at the Oneonta railyard. Since the 
department was informed that the fire was caused by gas, firemen left 
the station with the impression that they would encounter a tank car 
gasoline fire. 

The Oneonta Fire Department, the Schenevus Fire Department, and the 
Milford Fire Department, which are part of the 29 fire companies from 

- 8 - 

- 9 - 

Otsego County's Fire Department Mutual Aid program, responded to the 
alarm. They approached the accident via the north access road to the 
railroad's service road, passed the locomotive, and went directly to the 
accident . 

Water to fight the fire was supplied from tank trucks while hoses 
were being laid to the nearby Susquehanna River. Each fire company 
responded to the scene with two pumpers and a water tank truck. 

When firemen first arrived at the scene, they believed that they 
could control the fire and that it would be necessary to play water on 
the tank cars to keep them cool. This was the reason for laying hoses to 
the river. As each fire company arrived after Oneonta, they were in- 
structed to aid in this project. 

Just before the first explosion, firemen heard several tank-car 
safety valves operate and were ordered by two company chiefs to withdraw 
from the immediate area. Because there was no central communications 
system available, the withdrawal instruction probably did not reach each 
firefighter. Consequently, the area had not been completely evacuated 
when the explosion occurred, and a number of persons were injured. Most 
of the injured were firemen and several members of the press; the total 
injured was 54. No further attempts were made to cool or to extinguish 
the fire; instead, the area was secured, and injuries were treated. The 
fire was allowed to burn itself out. 

There was no pre-established system between the D&H railroad and the 
responding fire companies for the exchange of information either at the 
scene or by telephone contact. Few firemen knew that waybills describing 
the contents of each car were available on the train. Those who were 
aware of the waybills were not sure whether they were located on the en- 
gine or on the caboose. Usually, waybills are carried on the caboose, 
but on NWB-4, they were carried on the locomotive. Other than waybills, 
the D&H does not provide information on hazardous materials to the crew- 
members on either the locomotive or the caboose, except for the required 
notice of explosives. 

Paid firemen in the Oneonta Fire Department are required to take 
training courses, but the training of volunteer firemen is not mandatory. 
Most volunteer fire companies, however, require some training. Training 
courses and some training materials are provided by the New York State 
Fire Training Program, and a qualified State instructor teaches courses 
that are available to volunteer firemen. Training material is also re- 
ceived from the State of New York, the National Fire Protection Associa- 
tion (NFPA) , and the Bureau of Explosives of the Association of American 
Railroads (AAR) . 

The firemen were not familiar with CHEMTREC (The Manufacturing 
Chemists Association), and only a few knew of the Bureau of Explosives. 

- 10 - 

The chiefs of the volunteer departments believed that any information 
and advice regarding fighting unusual hazardous materials fires could be 
obtained from the county fire coordinator. No one knew of a specific 
point of contact with the D&H where they could obtain information con- 
cerning the trains or the railroad. 

Those firemen interviewed indicated they were not trained to deal 
with large volumes of LPG or to deal with a railroad accident. Their ex- 
posure to propane gas fires was limited to training information on 100- lb, 
home storage tanks. Most firemen had no concept of the force or potential 
of the quantity of the LPG carried by a single "jumbo" railroad tank car. 


The Accident 

The fractured center sill of Car 4 weakened the car's structure to 
the extent that the heavy lading caused the center sill to be deflected 
downward as the train proceeded toward the accident site. The center 
sill probably deflected after the train was dispatched from Binghamton. 
Such a deflection would not be detected during a running inspection, and 
crewmembers would not be able to detect the deflection during en route 

The downward deflection of the sill caused the north end of the sill, 
at the coupler, to be canted upward. Therefore, the pulling faces of the 
couplers were at an angle to each other, instead of parallel. The coupler 
of Car 4 was pulled upward as force was applied and eventually passed 
over the coupler of Car 3, and the train separated. 

The air hoses between Cars 3 and 4 were pulled apart when the 
train separated, which produced an automatic emergency application of 
the train brakes. Normally when an emergency application is made by a 
crewmember at the brake valve, it is readily discernible, and the en- 
gineer would have been aware if his action had initiated the emergency 

When the emergency brakes were applied on the locomotive, the power 
control switch reduced the power to idle. The emergency brakes on the 
locomotive and on the head 3 cars became effective before the brakes on 
the rear 118 cars. This difference in deceleration caused the rear of the 
train to collide with the slower-moving front portion, which produced an 
excessive run-in. Indications of such a collision were found on the 
couplers, the striking castings, and the ends of the head cars. 

The unusual movement felt by the crewmembers on the locomotive, which 
caused them to look rearward, was probably caused by the separation of the 
train. When the locomotive brakes were applied and the power was removed, 
the heavy axial force produced by the collision would have been applied 

- 11 

between the train and the third locomotive unit. Part of the longitudinal 
force would have been converted to a lateral force at the third locomotive 
unit. At the time of impact, the locomotive had entered the 3° 30' curve, 
which caused the locomotive and Car 1 to be at an angle to each other. 
The positions of the locomotive and car produced an outward force against 
the east, or high, rail. The force was greatest at that location, because 
of the resistive force available when the brakes were applied on the loco- 

The force caused the east rail to cant outward, which widened the 
gage of the track so that the west wheels of the third locomotive dropped 
inside the west rail. As the locomotive moved northward, the track was 
progressively destroyed by the wedging action of the wheels. 

When Car 4 collided with the front portion of the train it was prob- 
ably rocked violently, and the force of the collision turned it over 
along the west side of the No. 1 track. This was probably the reaction 
which the fireman and head breakman saw when they looked rearward. Cars 
5, 6, and 7 passed Car 4, derailed, and stopped along the west side of No. 
1 track. Cars 2 and 3 derailed and were dragged northward by the loco- 
motive. Cars 8 through 27 derailed and stopped on the track structure. 

Car 21, a tank car loaded with LPG , stopped crosswise on the track 
structure, and several cars overturned on its tank. The west portion of 
the tank was crushed and split open, which allowed a large volume of the 
LPG to escape. The escaping gas was ignited almost immediately by some 
source produced in the derailment. LPG probably flowed under the wreck- 
age, vaporized, and ignited, setting fire to the wreckage. Four of the 
derailed tank cars stopped in positions which permitted their tanks to 
be impinged by the fire. The tanks were heated, the operation of several 
safety valves was heard, and more fuel for the fire was provided. The 
fire increased in intensity, and tank sheets adjacent to the vapor spaces 
of the four exposed tanks heated to the point where the internal pressure 
in the tank exceeded the strength of the steel sheets, and the tanks 
ruptured . 

Need for Inspections 

The undetected failure in the center sill of the covered hopper sug- 
gests that cars should be inspected more thoroughly while they are in the 
shop. The failure was inside the hopper, and in normal inspections the 
car inspector does not see that part of the sill. Since the crack was 
progressive, the failure could have been found before the accident if 
the center sill had been accessible to inspectors. 

Since similar failures have not been reported and since the 15-per- 
cent sample of cars did not reveal such failures, the failure of the 
center sill on Car 4 was probably an isolated case. However, since the 

- 12 

failure occurred in an area that is not inspected regularly, owners of 
similarly constructed cars should check them more closely for defects. 

Track Structure 

The events of the derailment indicate that the east rail of the track 
was subjected to lateral forces that exceeded the rail's resistance. 
Since the evidence was not available to determine how many of the cross- 
ties in the curve were spiked with additional rail-holding spikes, the 
magnitude of the lateral and vertical forces, or the ratio or lateral to 
vertical forces required to overturn the rail, the track structure re- 
quired to resist overturning of the rail when the emergency brakes were 
applied could not be determined. 

There is reason to believe, however, that in this case the spiking 
and crossties required by Federal track safety standards would not have 
prevented the overturn, because the railroad was using three spikes on 
3° 30' curves when ties were replaced, and Federal Standards require only 
two. The railroad's experience had prompted them to use the additional 
spike. Therefore, although track in the derailment area met or exceeded 
Federal Track Safety Standards, it was not able to support the forces 
which resulted from an emergency brake application. Other Federal stand- 
ards require emergency braking under certain operating circumstances. 
Thus, two Federal Railroad Administration standards, one requiring emer- 
gency braking and the other requiring only two spikes per crosstie, are 

These standards should be changed so that they will be compatible. 
The Board discussed a general approach to the problem in its Special Study 
of Proposed Track Safety Standards, which states in part: 

"E. Future Development of Standards 

"The Safety Board recognizes the very difficult and complex task 
that FRA was faced with in these initial proposals, which were 
based as they must be on existing empirical approaches to defi- 
nition and standards, and the existing non-analytical track as- 
sembly methods. These constraints prevent the initial standards 
from resolving some of the problems which will develop in the future 
as a larger part of the field is covered. The Safety Board's ex- 
perience in reviewing the development of standards in other fields 
suggests that certain positive steps could be taken to smooth the 
advance of such standards: 

"1. Execution of the necessary testing and research to determine 
the actual degrees of various types of performance provided 
by track which meets the existing design-based standards. 
This would be a first step toward providing a full perform- 

- 13 - 

ance definition of the major functions of a railroad 

"2. Create uniform graded definitions for test purposes of 
various forms of environment, vehicles, operators, and 
system controls. The grading of track by speed, as in these 
regulations, is an example. There are as yet no uniform 
grades or ratings of various types of rolling stock or 
equipment operators. It is to be noted, for example, that 
these initial proposals do not provide any consideration 
of the various characteristics of rolling stock. The fact 
that track effects on rolling stock are not classified as 
yet prevents the inclusion of probable useful alternatives 
in these standards. All must be treated the same. 

"3. Determine, by study, all of the necessary functions of track 
in relation to other sub-systems of the railroad system and 
reclassify future standards according to functions. The 
present classes of roadbed, geometry, structure and appli- 
ances, if continued alone, will prevent the standards from 
being fully developed as performance standards. These 
classifications are useful for enforcement, but not defining 

"4. Be prepared to develop and employ both performance and design- 
based standards. It is quite practical to establish both 
forms of standards, the former to allow new technology and 
promote alternatives and the latter to provide simplicity in 
enforcement. When a given design has been qualified by test 
to prove that it fulfills the defined performance, its con- 
dition can be controlled by a design specification. In this 
way, a design specification and inspection format can be 
accepted for each different way of meeting the performance re- 
quirement. Both flexibility and enforcability can be obtained, 

"5. Reduce the range of variability in materials and assembly 
methods so that the results are more predictable. This 
variability not only requires unnecessarily high and wasteful 
safety factors, but insures that unknown hazards will exist 
when new technology such as very high speeds, welded rail, 
high center of gravity cars, and rapid deceleration rates 
are introduced. (See Safety Board report of accident at 
Crete, Nebraska, p. 43.)" 

This one accident occurring on a track which was believed to meet 
Federal track safety standards, strongly suggests that, with some combi- 
nations of locomotives and the location of loaded cars in the train, 
emergency braking which originates at the front of a train may consist- 
ently produce a derailment on such sharp curves by overturning the rail. 

14 - 

This particular problem inadequacy of track to sustain emergency braking, 
may not be the most frequantly encountered track problem, however, it 
does contain the elements of the general problem of avoiding derailment 
by physical failure of track under such braking conditions. The Board 
believes that an analytical approach to this problem will reveal the 
underlying problems of analysis and the need to consider long range staged 
development of track standards which will interrelate with equipment 
standards, as recommended earlier. A consideration of this problem will 
thus tend to show the need for a system-building concept including some of 
the considerations of Section E of the Safety Board's Special Study of 
Proposed Track Safety Standards . 

Tank Cars 

The tank cars involved in this accident were constructed similarly 
to those that have been involved in past catastrophic accidents. The 
safety problem was first addressed by the Safety Board in October 1969, 
in its report of an accident in Laurel, Mississippi. 1/ Following the 
Safety Board's recommendation, the Railway Progressive Institute (RPI) 
and the Association of American Railroads (AAR) established a research 
group titled the RPI-AAR Railroad Tank Car Safety Research and Test Proj- 
ect, to study the problem created when the tank cars are exposed to an 
accident or a fire. The study, with a few exceptions, has been completed, 
and several recommendations have been made to the industry and the FRA. 

0ns phase of the study involved full-scale fire tests of tanks 
loaded with propane. The tests are being conducted jointly with the FRA 
to determine if an insulating material satisfactory to tank car users 
could be developed for use on the tank car's exterior to lengthen the 
time a tank can be exposed to a fire before it ruptures. Although the 
tests have been completed, the data have not been evaluated to make rec- 
ommendations for its use. 

In this accident, a proper insulation would have lengthened the ex- 
posure time, and, perhaps, would have prevented the explosions by allow- 
ing heat to dissipate laterally from the hot area to colder parts of the 
tank. Under the circumstances, it could not be determined what amount of 
water would have been needed to prevent rupture indefinitely. 

Following the accidents at Laurel, Mississippi, and Crescent City, 
Illinois, 2^/ the rail industry and FRA became genuinely alarmed at the 
possibility of further catastrophe involving populated areas. To a con- 

JL/ NTSB Report of derailment of Southern Railway's freight train at 

Laurel, Mississippi, January 25, 1969. 
2/ Railroad Accident Report NTSB-RAR-72-2 , Toledo, Peoria, and Western 

Railroad Company's Train No. 20 at Crescent City, Illinois, June 21, 


- 15 - 

siderable degree, FRA has relied upon the voluntarily initiated research 
of Railway Progress Institute and Association of American Railroads to 
provide answers to the catastrophic fire threat when tank cars are in- 
volved in a rail crash. This field of effort was recommended to AAR by 
the Safety Board in October 1969, before FRA possessed comprehensive 
regulatory authority or a significant safety research budget of its own. 
The Safety Board reported to Congress on December 1, 1971, that another 
accident of this type (now represented by the Oneonta accident) could 
occur at any time, because the hazards had not been resolved. In October 
1972 , the Safety Board recommended that the tank car research be expe- 
dited. The Oneonta accident is the second major accident of this type 
since that advisory to Congress and the second which has produced more 
than 50 injuries to firemen. Nevertheless, it still appears that another 
accident of this nature could occur at any time and in a populated area. 

The ultimate responsibility to guard the bystanding public against 
the consequences of such accidents is that of FRA. The perspective in 
this task includes the fact that the smaller tank cars which preceded the 
"jumbo car" of specification 112A and 114A had a much lesser potential 
for catastrophic community damage by explosive ruptures because of the 
smaller quantity of fuel which could be released and because the smaller 
tank cars were insulated. A major cost-saving change in the larger cars 
was the omission of this insulation. As the Board has noted elsewhere, 
the type 112A cars were proliferated under Federal regulation without 
full-scale testing beforehand. The testing in public usage since that 
time has resulted in heavy loss to the public and firefighters. In fact, 
in four accidents involving heating and explosion of cars investigated by 
the Safety Board, the human loss has been borne exclusively by persons 
along the right-of-way and by firemen. There have been 3 fatalities and 
more than 300 injuries, which include 1 fatality and more than 140 in- 
juries to firemen. 

Given this accident experience, the period of testing in public 
usage should now be considered complete and ready for judgment. A proper 
finding from this period of usage is that movements of multiple car 
blocks of type 112A and 114A cars loaded with certain products can pro- 
duce community- size disasters. It is doubtful that these cars would have 
been permitted to come into use unaltered had their loss characteristics 
been known and had an opportunity to halt their proliferation been avail- 
able. However, the cars are now in use and an economy employing LPG fuel 
has developed around the availability of large shipments of LPG to areas 
distant from pipelines. The acceptably low cost of using this fuel in- 
cludes savings derived from omission of car insulation, use of much 
larger tanks, and shipment of nuiribers of loaded cars in a close-packed 
group to the same destination. 

One of FRA's tasks is that of converting the accident experience 
and research of the last 4% years into regulations that must reduce the 

- 16 - 

community hazard at least to the level which preceded the acceptance of 
the "junio tank car". The users of LPG have received benefits in terms 
of lower cost of fuel. Therefore, the task of FRA. is also one of deciding 
how much of this cost-saving to LPG users is to be employed for the pur- 
pose of protecting bystanders and firemen by requiring costly changes. 
The bystanders and firemen have been bearing the human cost of more ef- 
ficient LPG carriage. The acceptable cost of the corrections to be re- 
quired by regulation may be increased by another factor — tank car manu- 
facturers and users have already benefited economically by being permitted 
to use the lower cost, more efficient cars for long periods of time. 
About 15,000 cars have been permitted to be manufactured. The fact that 
the public has absorbed the human accident loss which made the cost savings 
possible should weigh strongly in the question of the cost of correcting 
new cars alone or, also, retrofitting the 15,000 existing cars. There is 
no way by which the existing tank cars can be made less sensitive to fire 
effects without retrofit, and they will constitute a large fraction of 
the cars in use for years ahead. 

Emergency Response by Firemen 

Evidence indicates that local firefighters may not be adequately 
trained or informed to fight fires involving 30,000-gallon tanks of LPG. 
The initial decision to cool the fire-impinged tanks was based on known 
procedures for management of propane fires; however, it appears that the 
fire authorities had insufficient information and expertise to support 
the on-scene decision regarding the larger tanks. Since the firefighters 
had been trained to fight fires involving only small amounts of propane, 
there was no way that the firefighters could determine the amount of water 
required to cool a 30,000-gallon tank. The lack of satisfactory training 
and materials for the local firemen was a definite factor in the injuries 
which resulted from the accident. 

The criticism in this analysis is not of the individual fire compan- 
ies or the dedicated firefighters who constitute them. The problem lies 
in the methods used by firefighters in the management of hazardous 
materials accidents, which continue to produce unnecessary casualties 
and losses. 

A lack of communications between firefighters was evident at the 
scene. When the fire increased in intensity, orders were issued for the 
firemen to pull back to a safer point; but before the orders could be 
executed, the tanks exploded. (See Figure 4.) 

Communications between the Fire Departments and the Railroads 

A railroad derailment or accident which involves tank cars loaded 
with hazardous materials can develop situations that are unusual for the 
average fire department. For this reason, railroads should provide the 

- 17 - 

- 18 - 

fire departments, in whose territory the railroad passes, with as much 
information as possible so that the firemen can properly assess the situa- 
tion and reduce the risks. The Safety Board's 1969 recommendations to 
the AAR along these lines has been acted upon by only one or two indi- 
vidual railroads, but not by the AAR. 

Fire departments should be instructed about waybills — what informa- 
tion is provided by the waybill, where they are carried on the train, 
which crewmember is in charge of the waybills, and how he can be identi- 

Information should be made available on both the locomotive and 
caboose concerning hazardous materials being carried by the train so that 
it will be readily available if an accident occurs. Firemen should have 
appropriate emergency telephone numbers beforehand so that further speci- 
fic information can be easily obtained at any time. 

In addition to being familiar with the information that can be pro- 
vided by the various transportation industries, firemen should also be 
trained regarding the techniques to be used in combating fires of the 
magnitude encountered in this accident. Such training could be accom- 
plished by the States, or by the proposed National Fire Academy at such 
time as it is established. Also, valuable information regarding the 
risks encountered in accidents involving the transportation of hazardous 
material by rail, which would be useful for training purposes, can be 
obtained from the Safety Board, the AAR and the FRA. 

After the derailment of the Missouri Pacific's train 94 and catas- 
trophic explosion of vinyl chloride at Houston, Texas, on October 19, 
1971, the Safety Board issued a safety letter ^/ to all State fire 
marshalls advising that, 

"In the absence of the information necessary for an evaluation 
of the risks of explosion or rocketing in a fire involving several 
tank cars, a prudent course of action may be the complete evacua- 
tion of the area within a radius of 2,000 feet. Exposure of emer- 
gency personnel to abrupt f lareups or explosions can be kept to the 
absolute minimum by this method when no other persons or minimal 
property are at risk. Under no circumstances should spectators be 
allowed in the potential danger radius of 2,000 feet. However, it 
is not possible to give all-inclusive advice concerning the degree 
of risk which should be taken to fight the fire in such accidents 
because there are so many variables involved. For example, there 
may be tank car fires in the vicinity of hospitals, schools, or 
other occupied premises which cannot be evacuated quickly." 

3/ NTSB Safety Recommendation R-71-38 issued December 27, 1971. 

- 19 - 

In the accident report 4/ the Safety Board also reconmended that: 

"The National Fire Protection Association, and similar standard- 
setting organizations, establish documented quality standards and 
quality control procedures for developing recommended practices to 
combat transportation emergencies which involve hazardous materials. 
These standards and procedures should include a requirement for a 
technical safety analysis of their applicability and the risks 
associated with their use. (Recommendation No. R- 72-45)" 

Although NFPA indicated that it was actively pursuing the goals sug- 
gested by the Safety Board's recommendations, there were no indications 
that the firefighters at this accident had been apprised of the relevant 
information. The New York Public Utility Commission in 1970 disseminated 
useful information on transportation fires to firefighters, but it ap- 
pears that it did not reach those firemen involved in this accident. 


1. D&H train NWB-4 was being operated according to company proce- 
dures as it entered the 3° 30' curve. 

2. No. 1 track in the area of the derailment probably met Federal 
track safety standards. 

3. Inspections made in Binghamton and in Oneonta and observations 
of the train en route did not detect the defect in Car 4, because the 
cracked center sill was not visible to inspectors. 

4. The fracture in the center sill started at a rivet hole and 
progressed across the vertical web of the sill. 

5. The center sill was deflected downward by weight and vibration 
in the vicinity of the crack which raised the coupler on the north end 
allowed it to pull over the coupler of Car 3. Consequently, the train 
separated . 

6. The emergency brake application following the train separation 
was the normally expected action under the provisions of 49 CFR 232. The 
force which overturned the east rail was the lateral component of a 
longitudinal compressive force in the train, generated by more effective 
emergency brake application at the locomotives. 

7. Whether two or three spikes were used to secure the rail to each 
tie on the 3° 30' curve could not be determined. 

4/ Railroad Accident Report NTSB-RAR-72-6, Derailment of Missouri 

Pacific Railroad Company's Train 94 at Houston, Texas, October 19, 

- 20 - 

8. The number of defective ties, if any, on track No. 1 in the 
3° 30' curve could not be determined. 

9. The east rail of No. 1 track was canted eastward as the loco- 
motive of NWB-4 moved northward around the curve, which resulted in wide 

10. The fire in the derailment resulted from the ignition by an un- 
known source, of escaping LPG from the crushed tank of the 21st car. 

11. The violent rupture of four tanks was a result of impingement of 
fire on uninsulated tank sheets which heated and weakened the steel. 

12. Regulations have not been promulgated by the FRA which would 
eliminate or reduce the severity of the problem which arises when tank 
cars loaded with hazardous materials are exposed to an accident or a fire. 

13. Recommendations have not been made by the RPI-AAR Railroad Tank 
Car Safety Research Test Project that would help to eliminate or reduce 
the severity of accidents and fires which involve tank cars loaded with 
hazardous materials. 

14. Firemen had not received information from the D & H on matters 
that would have aided them in combating the fire. 

15. Firemen had not been trained to combat fires of such magnitude. 
Further, they had not been trained for or exposed to information on the 
hazards that can result when tank cars loaded with hazardous materials 
are exposed to fire. 

16. The propulsion of large portions of tank cars 1,200 feet from the 
site was a result of rocketing action already displayed by type 112A and 
type 114A tank cars in previous accidents. 


The National Transportation Safety Board determines that the probable 
cause of this accident was the inability of the track to resist the 
lateral forces which canted the outside rail outward and widened the 
gage of the track. These forces which were induced at the third loco- 
motive unit resulted from the emergency application of the brakes when 
the train was separated between the third and fourth cars as it entered 
the 3° 30' curve. The train separated as a result of the broken center 
sill on the fourth car. 

- 21 - 

The National Transportation Safety Board recommends that: 

1. The Federal Railroad Administration (FRA) insure that inspection 
efforts of railroads to find defects on other cars having a center sill 
design similar to that which failed in this accident, are continued in 
periodic inspection until all such cars have been inspected. (Recommenda- 
tion R- 74-31) 

2. The FRA reevaluate the Federal Track Standards to determine 
whether the spiking and crosstie requirements for tracks on curves are 
adequate to prevent the rails from spreading when subjected to the lateral 
forces produced by an emergency application of train brakes from trains of 
a wide range of consists and operating speeds. (Recommendation R- 74-32) 

3. The FRA enploy the information developed by their tank car 
studies and by those of the RPI-AAR Railroad Tank Car Safety Research and 
Test Project and expedite the promulgation of regulations that will elim- 
inate or reduce to manageable dimensions the severity of the possible 
losses to communities which develop when tank cars of the general type of 
112A and 114A loaded with hazardous materials are exposed to a railroad 
accident or a fire. (Recommendation R- 74-33) 

4. The FRA promulgate regulations to require railroads to provide 
pertinent lading information to crewmembers on both the locomotive and 
caboose regarding those cars transporting hazardous materials. The in- 
formation could be presented in a manner similar to that required for ex- 
plosives in 49 CFR 174.589(f). In addition, these notices should contain 
the name of the commodity and its hazard rating. (Recommendation R- 74-34) 

The Safety Board reiterates and emphasizes the importance of the fol- 
lowing recommendations made in previous accident reports which have not 
been fully implemented and are applicable to this accident: 

Railroad Accident Report, Southern Railway Company, Laurel, Missis- 
sippi, January 25, 1969: 

"5. The Safety Board recommends that the Association of American 
Railroads and the American Short Line Railroad Association 
develop plans that will result in the fire chief of each com- 
munity through which the track of a member road passes knowing 
where immediate information can be obtained, describing the 
location and characteristics of all hazardous materials in any 
train involved in a train accident that affects a community. 
This recommendation can be accomplished in a relatively short 
time regardless of the level of training which may be achieved 
later by fire departments." (Recommendation No. R-69-22) 

- 22 - 

Railroad Accident report NTSB-RAR-70-2 , Illinois Central Railroad 
Company, Glendora, Mississippi, September 11, 1969: 

"2, The Federal Railroad Administration initiate research and 
development to provide prototype models of freight train 
braking systems 

(a) capable of providing shorter stopping distances which 
nearly approach the theoretical limits under all con- 
ditions of loading and length of trainsj 

(b) capable of stopping a train in the emergency applications 
now required by regulations without internal collisions, 
train separations, or damage to the train or its lading; 

(c) capable of propagating brake application, both service 
and emergency, throughout the length of train more ex- 
peditiously and surely; 

(d) capable of more rapid application of the full intended 
stopping force to the rails at each car after the appli- 
cation signal is received at each car." 

(Recommendation No. R-70-16) 












October 18, 1974 












Washington, D.C. 20591 

OfBcial Business 




DOT 513